While the display of high resolution magnetic resonance (MR) images in the operating room (OR) has revolutionized neuronavigation, tissue motion during surgery compromises the value of preoperative image information for navigational decision-making due to registration errors which accumulate between the dynamic OR space and the static preoperative image reference frame. While it is clear that intraoperative brain motion is significant, strategies to address this source of error during image-guidance are in early stages of development. A conceptually powerful approach to update preoperative images intraoperatively would be to generate a patient-specific computational model, collect readily accessible but incomplete intraoperative information relevant to brain motion with low-cost tracking technology, modify the model accordingly, and then update the preoperatively-obtained high resolution images. During the previous funding period, many of the elements needed to implement a modeling strategy in the OR have been developed and the confluence of these components has been demonstrated in a series of clinical cases analyzed retrospectively involving post-craniotomy brain sag which has been found to be one of the predominant modes of intraoperative brain deformation. The focus of this continuation application is two-fold: (i) to automate and merge the acquisition of incomplete intraoperative tissue motion data with an enhanced version of the existing computational model to provide updated images on a time-scale relevant to navigational decision-making directly in the OR, and (ii) to expand the use of preoperative image information in the construction of patient-specific models, namely to develop tissue property estimation schemes based on MR imaging. To achieve these goals, the specific aims are: (1) to extend model capabilities to include the later neurosurgical events of tissue retraction and resection and the mechanical response from viscoelastic and antisotropic behaviors, (2) to expand the acquisition of preoperative patient-specific image data to include mechanical and hydraulic property estimates based on MR elastography and diffusion tensor imaging, (3) to automate the intraoperative acquisition and integration of tissue motion data obtained from cortical surface scanning, coregistered ultrasound and instrumented retractors as model constraints, (4) to validate and assess in vivo the advances associated with Specific Aims 1-3 in the porcine brain, and (5) to evaluate model updating of preoperative images in human subjects using incomplete intraoperative data and full volumetric information obtained independently from intraoperative MR/CT. Successful completion of these aims is likely to lead to a methodology for retaining the registration accuracy of the high definition preoperative image of information which would preserve the value of image-guided procedures even in the face of considerable intraoperative brain motion.